fix: #247 logging for a sub-agent w/ stopAtToolNames (#531)

Author: seratch

.changeset/cuddly-states-begin.md

@@ -0,0 +1,5 @@
+---
+'@openai/agents-core': patch
+---
+
+fix: #247 logging for a sub-agent w/ stopAtToolNames

packages/agents-core/src/agent.ts

@@ -556,6 +556,20 @@ export class Agent<
  context,
  ...(runOptions ?? {}),
  });
+
+ const usesStopAtToolNames =
+ typeof this.toolUseBehavior === 'object' &&
+ this.toolUseBehavior !== null &&
+ 'stopAtToolNames' in this.toolUseBehavior;
+
+ if (
+ typeof customOutputExtractor !== 'function' &&
+ usesStopAtToolNames
+ ) {
+ logger.debug(
+ `You're passing the agent (name: ${this.name}) with toolUseBehavior.stopAtToolNames configured as a tool to a different agent; this may not work as you expect. You may want to have a wrapper function tool to consistently return the final output.`,
+ );
+ }
  const outputText =
  typeof customOutputExtractor === 'function'
  ? await customOutputExtractor(result as any)

packages/agents-core/test/agent.test.ts

@@ -7,6 +7,7 @@ import { z } from 'zod';
 import { JsonSchemaDefinition, setDefaultModelProvider } from '../src';
 import { FakeModelProvider } from './stubs';
 import { Runner, RunConfig } from '../src/run';
+import logger from '../src/logger';
 
 describe('Agent', () => {
  afterEach(() => {
@@ -186,6 +187,30 @@ describe('Agent', () => {
  expect(result2).toBe('Hello World');
  });
 
+ it('warns when using asTool with stopAtToolNames behavior without custom extractor', async () => {
+ const warnSpy = vi.spyOn(logger, 'debug').mockImplementation(() => {});
+ const runSpy = vi.spyOn(Runner.prototype, 'run').mockResolvedValue({
+ rawResponses: [{ output: [] }],
+ } as any);
+
+ const agent = new Agent({
+ name: 'Stopper Agent',
+ instructions: 'Stop instructions.',
+ toolUseBehavior: { stopAtToolNames: ['report'] },
+ });
+
+ const tool = agent.asTool({
+ toolDescription: 'desc',
+ });
+
+ await tool.invoke(new RunContext(), '{"input":"value"}');
+
+ expect(runSpy).toHaveBeenCalledTimes(1);
+ expect(warnSpy).toHaveBeenCalledWith(
+ `You're passing the agent (name: Stopper Agent) with toolUseBehavior.stopAtToolNames configured as a tool to a different agent; this may not work as you expect. You may want to have a wrapper function tool to consistently return the final output.`,
+ );
+ });
+
  it('allows configuring needsApproval when using an agent as a tool', async () => {
  const approval = vi.fn().mockResolvedValue(true);
  const agent = new Agent({